Statistical Learning: The role of implicit and explicit processes in sequential regularities

Menchinelli, Federica (2019) Statistical Learning: The role of implicit and explicit processes in sequential regularities. PhD thesis, University of Lincoln.

Statistical Learning: The role of implicit and explicit processes in sequential regularities
Menchinelli, Frederica - PhD - Psychology.pdf - Whole Document

Item Type:Thesis (PhD)
Item Status:Live Archive


Within cognitive psychology, Statistical Learning (SL) refers to our use of the statistical information available in our sensory environment to extract relationships between stimuli which unfold over time. SL enables us to use previous and current events to make predictions on upcoming ones, and it is at the basis of a number of cognitive functions (Bertels, Boursain, Destrebecqz, & Gaillard, 2015a). Aiming to address the lack of systematic investigations in the area, this thesis is concerned with the type of knowledge which results from auditory and visual SL, and whether it is implicit, unconscious, or explicit, conscious. We aimed to address existing methodological challenges around the measurement of conscious knowledge through the adoption of a novel use of the Process Dissociation Procedure in the context of a forced-choice task, in combination with the guessing and zero correlation criteria. Chapter 2 established the use of measures to assess the status of conscious knowledge of auditory stimuli generated through a transition matrix. We found successful learning of these stimuli in both adults and children, and that both age groups develop an awareness of the knowledge that they had acquired. In Chapter 3 we studied knowledge status in a triplet learning paradigm in both the auditory and visual modality. Our measures indicated that participants were fully aware of the visual and auditory stimuli learned. Chapter 4 was aimed at validating and consolidating our findings of explicit knowledge by using a combination of direct and indirect measures in a visual triplet learning paradigm, and additionally compared adults and children. The knowledge acquired was prevalently explicit in both age groups. We also found that participants’ awareness of the acquired knowledge did not coincide with the ability to reproduce the training material in a generation task. Chapter 5 investigated the hypothesis that implicit and explicit knowledge are dependent on the speed of stimulus presentation. We found that, although statistical learning can take place at different presentation speeds, participants’ knowledge is weaker at a faster speed. Knowledge appeared more implicit at faster stimulus presentation speeds and more explicit at slower speeds. In Chapter 6 we investigated the electrophysiological correlates of implicit and explicit knowledge in visual statistical learning. There were no differences in in ERPs between implicitly and explicitly-learned stimuli in either learners or non-learners within our sample. However, we found suggestions that a learning effect may be present and detectable through the ERPs in the absence of above-chance behavioural performance. This PhD builds on, and extends, the existing literature, and sheds light on the theoretical and methodological challenges inherent to the behavioural approach in statistical learning. We put forward the hypothesis that knowledge measured behaviourally tends to become more explicit, the greater the learning effect, and that contradictions between measures of conscious knowledge arise in the presence of low learning. We explore promising approaches for future research to advance knowledge about statistical learning and the type of knowledge acquired, and we make a case for the use of combined electrophysiological and behavioural methods.

Divisions:College of Social Science > School of Psychology
ID Code:44796
Deposited On:05 May 2021 10:45

Repository Staff Only: item control page